26 Proceedings of the Royal Society 
The apparatus which we used for the preliminary experiments is 
shown in diagrammatic section in the figure. It consisted of a cylin- 
drical gas reservoir (A, A) of tinned iron, about 4 feet in length 
and 4 '5 inches in diameter. This reservoir may be • called the 
absorber. It was placed with its axis vertical. It had three 
openings B, C, D, of which the two (B and C) were fitted with 
metal tubes and stopcocks, and were used for filling the reservoir 
with the gas under investigation, and the third (D) was fitted with 
a metal tube holding the manometer E. The curved surface and 
the bottom of the absorber were surrounded by another cylindrical 
vessel F, F, also of tinned iron, whose sides were everywhere 
about 0-5 inch from the sides of the absorber. This outer casing 
had two openings (G, H) fitted With tubes and stopcocks, the one 
at the bottom, the other at the top. By their means the space 
between A and F could be kept full of running water. On the 
upper end of the absorber was a flat cylindrical box (K) about 
1 inch in depth. The upper end of the cylinder formed the bottom 
of the box. This box, which may be called the radiator, was also 
provided with two stopcocks and tubes L, M, placed as in the 
diagram, by means of which it could be filled with either water or 
steam. For this purpose an india-rubber tube from M dipped under 
water in a sink, while B could be put in communication by means 
of a three-way tube with either a boiler or the water supply. The 
absorber was thus completely surrounded by a jacket divided into 
two compartments, one of which was the radiator. The manometer 
E consisted of a bent tube of about 1“”" bore. It was provided 
with a stopcock N, was fitted in the tube D by means of an india- 
rubber stopper, and contained dilute sulphuric acid of known 
density. The levels of the liquid were read off by the aid of a 
card, with a divided scale, which was attached to the tube. 
The following is a description of 'our mode of observation. The 
absorber was first filled with the gas under investigation. Thus, 
to fill with dry air, B was attached to a series of drying tubes 
(sulphuric acid and chloride of calcium) and C to the suction pump, 
and a current was kept up until the dry air had displaced the damp 
air which at first filled A. To fill with moist air, the drying tubes 
were replaced by wetting tubes. To fill with olefiant gas, C was 
left open, and the gas driven in from the generator at B until the 
